Comment by smokel
4 days ago
> Frivolous use of the term World Model
The term "world model" as it was once used in model-based RL can now apparently refer to anything as silly as linear regression. Then again, the RL folks probably borrowed the term from behavioral scientists before them. It's probably best to simply accept this :/
A similar thing happened to "object oriented" which has been misused by philosophers and visual artists alike.
I'd have way more sympathy for the world-model-term-misuse complaints if the ML world hadn't studiously ignored and then reinvented so many fields over the years.
https://arxiv.org/abs/1805.06485 is enough to make any half decent game dev sit around wondering what was going on, as well as the general "maybe we're wasting a lot of space with all these floats?" At least now LLMs can tell them what SoTA is in other fields before they try to rediscover it.
> probably borrowed the term from behavioral scientists
It's from epistemology. It is not there to refer to the subjective but to the objective.
> A similar thing happened to "object oriented" which has been misused by philosophers and visual artists alike
For instance?
In the 1990s "Object-Oriented Ontology" was introduced by Graham Harman [1]. The name was borrowed from computing, but its meaning has little or nothing to do with Simula or Smalltalk.
[1] https://en.wikipedia.org/wiki/Graham_Harman
This is very interesting especially in the terms of what I think some call a "rabbit hole",
but we could be curious on how and why you saw misuse.
The OOO/"Speculative Realism" crowd are loony charlatans who make the current post-modern neo marxist politburo who runs the humanities parts of academia look downright sane.
My response to unironic OOO believers is that we must wage war on objects:
https://en.wikipedia.org/wiki/Resistentialism
I don’t see a problem. It appears to use a shared latent to learn inverse dynamics.